Research On Active Power Measurement Of The Optimal Initial Angle Sampling And Wavelet Packet

Digital sampling power (electrical energy) measurement technology is widelyused in the field of electrical automation. It provide a technique guide for promotingthe development of electrical measuring instrument and maintain the interests of powergeneration, power supply and the users. At present，the common power measurementtechnology exists some limitations in practice, which is mainly manifested in thecontradiction between measurement precision and speed. At the same time, under thenew situation of smart grid, new features of grid signal, such as steady-state periodicsignal contained harmonic, transient signal etc. urge us to do some in-depth,comprehensive and systematic research on electric power measurement error ofAC-sampling, also inspire us to explore a new sampling method to meet the accuracyand real-time performance of electrical power measurement.The thesis focus on solving the problem of steady state signal active measurementprecision and speed and to explore the unsteady signal active measurement method,mainly carried out the following work:Firstly, the basic principle of synchronous sampling, quasi-synchronous samplingand asynchronous sampling are elaborated in this thesis.The causes of error of activepower measurement under different sampling mode are summarized. And theimprovement methods currently used are evaluated systematically. Those researcheslay the foundation for the study of active power measurement error and methods underasynchronous sampling.Secondly, this paper analyzes the error of active power measurement insingle-phase sinusoidal and single-phase non-sinusoidal systems under asynchronoussampling, and puts forward a kind of high accuracy active power measurement methodbased on optimal sampling initial angle-the optimal sampling method. Then the effectof sampling number on power measurement error under the optimal sampling methodis discussed. From engineering point of view，deriving balanced three-phase sinusoidal,three-phase asymmetric sinusoidal system error expressions of the total active power，to explain that this method is also applicable to three-phase system when the system isstable. Simulation result show that the optimal sampling method can effectivelyimprove active power measurement precision, the measuring accuracy is higher than the conventional zero point sampling method, especially for sinusoidal system, themeasurement precision can be improved one order of magnitude.Thirdly, for the optimal sampling method, the paper designs an active real-timemeasurement system including voltage and current transformation, optimal samplingtrigger pulse generating circuit, data acquisition and processing unit. And speciallydesigns a phase shifter controlled by power factor to achieve sampling trigger signal.Multisim simulation result shows that the accuracy of the phase shifter can reach0.1°within the scope of the phase shift(ranges from0°to180°), and the response speed isfast(only needs a period), the corresponding active measurement theoretic error isaround0.03%. The circuit design is simple and has a good real-time performance.Thismethod overcomes the disadvantage of the lack of standardization of softwareprogramming and a bad real-time performance, can apply to ADI company chipintegration, simplify the software and hardware design of power measurement.Finally, explores the effectiveness of the active power calculation using thewavelet packet method and its adaptability to the actual signal. The selecting principlesof wavelet basis are discussed, and points out that the quality of signal analysis bywavelet packet is strongly influenced by the selected wavelet basis. MATLABsimulation result shows that the calculation precision of active power can reach10-4orders of magnitude when choosing the suitable wavelet basis and the method has astrong adaptability to the actual signal. Analysis and discusses the influence offrequency fluctuations on the method of active calculation accuracy. It shows that themethod of active power calculation accuracy will be affected by non-integer periodsampling, and the improved method remains to be further research and exploration.